Deep Learning VM Image

[EXAMPTOPIC] Deep Learning VM Image : 딥러닝 모델 학습에 필요한 모든 구성요소가 갖춰져있는 VM 관련 내용을 정리합니다.

Deep Learning VM Image

• Properties of Deep Learning VM Image
• [EXAMTOPIC] 3 Trouble Shooting cases for Deep Learning VM Image
  (1) Quota exceeded, (2) Resource not found, (3) Preemptible instances
• [EXAMTOPIC] Q 52

Deep Learning VM Image

Deep Learning VM Image : 구글클라우드 엔진에서 딥러닝 모델 학습에 필요한 모든 구성요소가 갖춰져있는 VM

• Preconfigured virtual machine images optimized for DS/ML tasks ⇒ build deep learning projects fast on a Google Compute Engine instance.

Key Feautres

• automate your notebook training ⇒ No need to pay extra or manually manage Cloud infrastructure
• launch Compute Engine instances pre-installed with TensorFlow, PyTorch, scikit-learn and more and tested for compatibility
• Fast prototyping
• Accelerate data processing tasks : add Cloud GPU and Cloud TPU support
• Support many combinations of framework and processor : TensorFlow Enterprise, TensorFlow, PyTorch, and generic high-performance computing with versions for both CPU-only and GPU-enabled workflows
• Optimized for performance (accelerate model training and deployment with the latest NVIDIA® CUDA-X AI libraries and drivers and the Intel® Math Kernel Library)
• Integrated support for JupyerLab

Preemptible VMs 선점형 가상머신

• Low-cost.
• Short-duration VM option for batch jobs and fault-tolerant workloads.
• Can terminate at anytime and will terminate within 24 hours.
• Cannot migrate to regular VM and no SLA.

Resource & Articles

• Let Deep Learning VMs and Jupyter notebooks to burn the midnight oil for you: robust and automated training with Papermill | Google Cloud Blog
• Deep Learning VM Image| Google Cloud
• Use Deep Learning VM Images and Deep Learning Containers with Vertex AI
• All things GCP: Machine Learning Decision pyramid | LaptrinhX

EXAMTOPIC - 3 Trouble Shooting cases for Deep Learning VM Image

Quota exceeded, Resource not found, Preemptible instances

1. Quota exceeded 할당량초과

- Quota 'NVIDIA_K80_GPUS' exceeded. Limit: 0.0 in region us-east1. : You do not have enough quota.

• Solution : Check the quotas page to ensure that you have enough GPUs available in your project ⇒ If GPUs are not listed on the quotas page or you require additional GPU quota, request a quota increase.
  • preemptible GPUs & normal GPUs require separate quota requests. You can't use preemptible GPU quota for normal GPUs.
    • quota is per region ⇒ creating the instance in the region where you have quota.

2. Resource not found

- The resource 'projects/deeplearning-platform/zones/europe-west4-c/acceleratorTypes/nvidia-tesla-k80' was not found
: You are trying to create an instance with one or more GPUs in a region where GPUs are not available (for example, an instance with a K80 GPU in europe-west4-c).

• Solution: To determine which region has the required GPU, see GPUs on Compute Engine.

3. Preemptible instances

I can't create preemptible instance from the UI, even though I have quota. 할당량이 있지만, UI에서 선점형 인스턴스를 만들 수 없는 문제

• Solution : preemptible instance can't be created from Google Cloud Marketplace ⇒ must use the CLI. ⇒ add --preemptible when setting up new instance.

EXAMTOPIC Q 52.

You are training a deep learning model for semantic image segmentation with reduced training time. While using a Deep Learning VM Image, you receive the following error: The resource 'projects/deeplearning-platforn/zones/europe-west4-c/acceleratorTypes/nvidia-tesla-k80' was not found. What should you do?

• ❌ A. Ensure that you have GPU quota in the selected region.
• ⭕ B. Ensure that the required GPU is available in the selected region.
• ❌ C. Ensure that you have preemptible GPU quota in the selected region.
• ❌ D. Ensure that the selected GPU has enough GPU memory for the workload.